Abstract: Systems and methods that enable high quality audio teleconferencing are disclosed. In one embodiment of the present invention, a signal processor receives signals from a spatially dispersed set of directional microphones, processing the microphone signals and the far-end received audio into a signal for transmission to a far-end party. The processing may comprise the use of one or more algorithms that reduce conference room noise and may selectively increase participant audio levels by processing the microphone signals using beamforming techniques. An embodiment of the present invention may also comprise one or more omni-directional microphones that may be used in cooperation with the directional microphones to adjust for background noise, acoustic echo, and the existence of side conversations.

Abstract: Apparatuses, systems, and methods for a multi-carrier communication system that detects and reduces an echo from a non-linear element present on the transmission medium. In an embodiment, a training period is established between a first transmitter-receiver device and a second transmitter-receiver device in the discrete multiple tone system that separates communication signals into two or more separate frequency bands. Noise caused by an echo generated by a non-linear element present on the transmission medium is detected during the training period. The significance of the non-linear echo contribution to the overall ambient noise level present in the system may be determined.

Abstract: A noise level calculator for detecting noise in a telephone line echo canceller that utilizes adaptive filters enables silencing of noise switching so that it is not heard on the far end side. The calculator promotes determining whether the reference signal applied to the adaptive filter is a noise or non-noise segment. The calculator operates by calculating the difference in energy of two sample windows, calculating the difference in the energy of the signal within each of the sample windows, updating a variance parameter based on the difference, and performing further alternative operations based on the value of the variance parameter.

Abstract: A system for updating filter coefficients for use with an echo canceller, a related method for updating filter coefficients and an echo canceller including the system or the method. In one embodiment, the echo canceller has a nonadaptive filter and an adaptive filter, the nonadaptive filter having nonadaptive filter coefficients and configured to develop a nonadaptive error signal as a function of the nonadaptive filter coefficients, and the adaptive filter having adaptive filter coefficients and configured to develop an adaptive error signal as a function of the adaptive filter coefficients. The system includes a monitoring subsystem configured to monitor a quality of the nonadaptive and adaptive error signals and a selection subsystem configured to select one of the nonadaptive and adaptive filter coefficients for use by the nonadaptive filter based on the quality of the nonadaptive and adaptive error signals.

Abstract: An adaptive filter is programmed with an algorithm based on a normalized Least Mean Squares (nLMS) algorithm that adapts each sample time. The algorithm is modified to be more efficient in a variety of DSPs by computing multiple errors, one per sample, before updating coefficients. The update equation utilizes the multiple errors to achieve adaptation at a similar performance to known nLMS algorithms that adapt each sample time but without the instability that is observed in low echo-to-near-end-noise ratio (ENR) input conditions. Varying the relaxation step size prevents divergence. The DSP utilizes either one or more MAC units.

Abstract: The present invention provides for adaptive filters that have improved computational and memory bandwidth proprieties. When applied to telecommunication applications, the present invention additionally provides for improved methods and systems of canceling echoes. In one embodiment of the adaptive filter of the present invention, a filter, preferably an adaptive finite impulse response (FIR) filter, of an appropriate length, N, is chosen. Once the filter is chosen, convergence is achieved and the filter is converted to an infinite impulse response (IIR) filter. In the course of operation, data is received from an input source and used to adapt the zeroes of the IIR filter using the least means square (LMS) approach, keeping the poles fixed. The adaptation process generates a set of converged filter coefficients that are then applied to the input signal to create a modified signal used to filter the data.

Abstract: An apparatus and technique for determining whether a telephone loop is capable of carrying a Digital Subscriber Line (DSL) signal. Time domain reflected (TDR) signals that are echoes of transmit impulse signals are processed using time domain echo averaging and echo enhancement using second-order statistics. Each technique allows clearer visualization of the TDR echo signal by reducing the noise and distortion present in the TDR signals. These statistical signal processing techniques allow determination of telephone loop characteristics such as loop length once the TDR echo signal amplitude and time index are calculated.

Abstract: The present invention provides an acoustic echo canceler for use in a telephone set having a handsfree conversation function. This acoustic echo canceler simultaneously achieves both the stabilization and speedup of an estimation operation and also attains the more improvement of speech quality. Moreover, the present invention provides a handsfree telephone set employing this acoustic echo canceler. Furthermore, this acoustic echo canceler is operative to cancel an echo generated by inputting an output of an acoustic device to itself substantially simultaneously with the outputting the output thereof. Further, this acoustic echo canceler comprises an adaptive filter for performing an operation of estimating the frequency characteristic and the phase characteristic of a part, whose frequency characteristic and phase characteristic vary, on an acoustic path and a fixed coefficient filter for simulating a part, whose frequency characteristic and phase characteristic are fixed, on the acoustic path.

Abstract: A method of controlling an echo canceller in a communications channel, is disclosed wherein input signals from the communications channel are first subbanded into a subband. Echo locations are then identified within the subband and are used to control the echo canceller. Typically, the echo canceller will be a fullband echo canceller with an adaptive filter, in which case the echo locations are used to control the filter coefficients.

Abstract: A method of converging a step size control for an adaptive filter of a communication channel including: (1) initializing a nominal step size value and a penalty point value; (2) combining the nominal step size value and the penalty point value to generate a step size value; and (3) dynamically changing the step size value in response to a characteristic measure of a quality of the communication channel. The step size value is changed by adjusting the nominal step size value, the penalty point value, or both. In a preferred embodiment the penalty point value is adjusted dependant on: (1) a tone originating from the far end (2) full convergence (3) the power level of a residual error signal (4) the channel's near-end background noise and/or (5) weak double-talk in the communication channel. The nominal step size value is adjusted when an achieved initial combined loss is about 15 dB or greater and is reset by divergence.

Abstract: An echo canceller, which includes, receives and sends paths connected to standard telecommunication interfaces, a non adaptive filter with filter reflection coefficients which generate an estimated echo signal, which when subtracted from the send path input produces a nearly echo free send path output, an adaptive filter generating a second estimated echo signal, which when subtracted from the send signal, provides a possible alternative nearly echo free send path signal, a fast non iterative least squares method of estimating a reflection, a controller means which, in response to an observed non echo free send path output, causes the fast non iterative least squares method to search for possible new reflections, and a controller means for expurgating unnecessary reflection coefficients in the non adaptive filter.

Abstract: Device and method for noise suppression in telephone equipment, in particular with hands-free equipment, having a level discriminator, an additional filter with a group propagation time of less than 2 ms being inserted into the transmission and receiving paths of the telephone equipment.

Abstract: A communication system having an echo canceller is disclosed. One embodiment of the echo canceller includes an adaptive filter used to provide an estimate of reflected echo which is removed from the send signal. The echo canceller may also include a near-end talker signal detector which may be used to prevent the adaptive filter from adapting when a near-end talker signal is present. The echo canceller may also include a nonlinear processor used to further reduce any residual echo and to preserve background noise. The echo canceller may also include a monitor and control unit which may be used to monitor the filter coefficients and gain of the adaptive filter to maintain stability of the echo canceller, estimate pure delay, detect a tone, and inject a training signal. The echo canceller may also include a nonadaptive filter used to reduce the length of the adaptive filter.

Abstract: A method of estimating an echo return loss of a communication link measures the peak power value of a signal received from a far end of a communication link and the corresponding peak power value of a signal received from the near end of the communication link. When the near end signal is free from near end voice and excessive noise, the near end corresponding peak will be the echo of the far end power. The far end power is monitored and when a peak is detected, a period, which can be equal to a filter length period, is initiated. When the period has run, the delay line attached to the near end is polled to determine the peak power in the delay line. This peak power will correspond to the far end peak. If a peak, exceeding the first peak is detected before the period has run, the period is reinitiated. The echo return loss estimate is calculated as a ratio of the far-end peak power value to the near-end peak power value.

Abstract: An echo canceller includes two parallel filters and a controller. The controller chooses the best of the two filters for the final echo cancellation. The filter coefficients are exchanged between the two filters constantly so that both filters retain performance all the time even if double-talk and echo path variations occur very closely in time. One filter is strongly robust and will not diverge during double-talk, and another filter is weakly robust and will converge rapidly during echo path variation.

Abstract: A communication system being made up of a two-wire/four-wire hybrid circuit, an echo canceller, and a control device and having functions of forming a plurality of communication paths and changing communication paths is provided which is capable of ensuring high speech quality by suppressing and canceling an echo occurring in telephone speech after establishment of path connection. The communication system also includes a training signal source to feed a training signal to the echo canceller, a switching circuit to do switching of a signal to be fed to the echo canceller between an ordinary signal and the training signal, and a control device to exert control on path connection and path changing and to feed the training signal to the switching circuit to make a training signal be fed from the training signal source to the echo canceller and to have the echo canceller converge its training operation.

Abstract: An echo cancelling system for cancelling echoes in a communication path includes an echo locator and an echo canceller. The echo locator is coupled to the communication path and locates the positions of echo signals received from the communication path. The echo canceller is also coupled to the communication path and cancels echo signals received therefrom. During echo cancellation, the echo canceller becomes active in regions corresponding to the positions of echo signals in response to the echo locator.

Abstract: The invention provides an apparatus and method for bulk delay insertion prior to echo cancellation, with echo suppression. The exemplary apparatus embodiment (100) bandpass filters (120) and down samples (130), by a down sampling factor, a reference signal and a combined signal that includes an echo of the reference signal. An adaptive echo delay filter (140) determines a plurality of filter coefficients, using the down sampled reference signal, and using an error signal formed as a difference between the down sampled combined signal and the adaptive echo delay filter output. When the filter coefficients have converged, an echo delay time period is determined as a product of the tap location of the maximum filter coefficient and the down sampling factor. When the echo delay time period is greater than the capacity of an associated echo canceller, an adaptive delay inserter (110) inserts the echo delay time period prior to echo cancellation.

Abstract: A process is described which generates matrix coefficients using zero-lag auto and cross-correlations from signals commonly found in echo cancellers. Double-talk and path changes are then detected using matrix operations such as determinants, eigendecompositions, or singular value decompositions (SVDs). In a preferred embodiment, the determinant of the correlation-based matrix is compared against predetermined threshold values.

Abstract: An echo canceler for generating an echo replica and subtracting the echo replica from a local input signal to create a residual signal for outgoing transmission. An echo canceler has adaptive filters that have different computational accuracy generating the echo replica. An evaluator coupled to the adaptive filters groups coefficients into each of segments, and evaluating a computational accuracy information according to the coefficients for the segments. A computational accuracy controller in the evaluator assigns the local input signal to the adaptive filters, and adjusts the coefficients according to the computational accuracy information.

Abstract: A communication transmitting and receiving system in which the effects of near-end echo and near-end crosstalk signal from the communication medium are mitigated by adaptively reproducing the near-end echo and near-end crosstalk signal, which is then subtracted from the received signal. Filter coefficients for a Finite Impulse Response filter are adaptively generated to reproduce the near-end echo and near-end crosstalk. The filter coefficients are regenerated for the Finite Impulse Response filter in an adaptive correlator at the arrival of each received signal and whereby each new filter coefficient is a weighted sum of a previous coefficient and one received signal multiplied by a time delayed version of one transmitted signal.

Abstract: Apparatus and accompanying methods for use in a long-haul echo control device to automatically measure sufficient changes in path length and compensate a long-haul echo canceller for resulting variations in echo occurring over a path. The apparatus measures a round-trip path delay interval of, e.g., a predefined signaling tone(s) sent between international switching centers while establishing an outbound call connection over the path. Once the delay is measured, the apparatus sets echo delay of a corresponding long-haul echo canceller to accommodate that delay. For a multi-channel facility (e.g., T1 or E1), path delay is separately measured for each channel on that facility, and then the apparatus changes the echo delay to the same amount for the corresponding echo canceller associated with each and every channel on that facility.

Abstract: Long distance telephone connections must commonly employ echo cancellers to eliminate irritating echo. The normal practice is to additionally employ non-linear processors (NLPs) as an integral component of an echo canceller in order to eliminate small residual echo which remains after initial echo cancellation. While such devices do indeed remove remaining residual echo, they also may introduce additional problems, such as suppression of the background signal. However, in some cases there is no need to deploy a potentially problematic non-linear processor. If the overall route delay of the telephone connection is not excessive, the participants do not notice the residual echo. Therefore, a determination is made of the end to end delay of the overall telephone connection. If the determined delay is not excessive, non-linear processing is not applied or is applied less aggressively. This results in a clearer telephone call, free from non-linear processor problems.

Abstract: An adaptive digital filter (10) is adapted to generate a pseudo echo Se which is produced by imitating a noise, using part of receiving signals Si and to output it. An echo produced through an echo path (15) is mixed, as an echo noise, with a sending signal So and this sending signal is attenuated by a level attenuator (6). The level attenuator (6), in order to match the amplitude level of the echo noise of the sending signal So to that of the pseudo echo, attenuates the sending signal So and feeds it to a subtracter (7). The subtracter (7) eliminates echo noises by subtracting the pseudo echo Se from the sending signal So. Even if the echo path is in the state of amplified system, the complete elimination of the echo can be achieved accordingly.

Abstract: A digitized input signal is sampled, to obtain a reference input signal. The reference input signal is spuriously whitened, to obtain a pseudo whitened signal. An echo replica signal is generated on the basis of the pseudo whitened signal and an adaptive filter coefficient. A cancellation error signal is obtained on the basis of an echo signal and the echo replica signal. The adaptive filter coefficient is updated on the basis of the cancellation error signal, the adaptive filter coefficient, and the pseudo whitened signal.

Abstract: Real and imaginary components of the discrete Fourier transform (DFT) of the residual echo in an acoustic echo cancellation system are separately estimated using mapping functions derived without an assumption of statistical independence between separate components of a near-end microphone input signal. The residual echo estimator calculates the real and imaginary components using an approximation of the microphone input signal without the echo. A different approximation is used depending on whether the near-end talker and/or the far-end talker are talking.

Abstract: The present invention provides a voice input apparatus which can extract an input voice without interrupting a generated background sound. When a talk switch 50 is pushed down, a LMS algorithm processing part 26 stops an update operation of a filter coefficient of an adaptive filter 24 and stores filter coefficient W1 at this moment in a portion for storing filter coefficients 30. A portion for calculating filter coefficients 44 calculates a filter coefficient W2 of a filter 42 to simulate a sound-transmittal line of an audio sound which is produced from a speaker 102 and is input into microphones 10, 12 by means of these stored filter coefficient W1 and various transmittal characteristics CS1, CS2 and the like. By extracting an output signal Y4 of the filter 42 from an output signal Y3 of a portion for eliminating surrounding noise 20 by means of a computing part 46, audio sound components included with an operating voice collected by the microphones 10, 12 are removed.

Abstract: A method and apparatus that performs specific automatic gain control in an echo canceller (16). The disclosed method and apparatus utilize a low complexity algorithm that performs automatic gain control to achieve greater speed of convergence of the echo canceller (16) and an increased ability to distinguish between near-end speech and echo through the use of specific automatic gain control. A double talk threshold of the echo canceller is adjusted by using, in part, computed gains of inbound and outbound signals internally within the echo canceller in order to adjust the echo canceller double talk threshold.

Abstract: A method and apparatus is capable of detecting the presence of a non-linear echo path as well as transitions from non-linear echo path to linear echo path and vice versa. Additionally, when non-linear echo path is detected, the echo canceller (12) is configured to follow a prescribed logic routine that combats possible echo leak-through as a result of compromised cancellation due to non-linearity. This prescribed logic routine affords the echo canceller (12) protection against non-linear echo path by preserving an already converged state of the echo canceller (12) from possible divergence due to the occurrence of non-linearity in the echo path and thereby prevents echo leak-through that would occur due to divergence of the echo canceller (12).

Abstract: A base band echo cancellation method and apparatus utilizing Laguerre echo estimation. The method begins with a series of transmitted symbols that are stored in a buffer. The buffer is segmented into a head portion and a tail portion. The symbols in the tail portion are passed through a Laguerre or IIR filter and then reinserted into the tail portion of the buffer. An echo estimation signal is generated based on the symbols in the buffer. The apparatus includes a transmit symbol buffer that is coupled through a modulator to a Laguerre or IIR filter. A demodulator couples the filter back to the buffer. An echo estimator is also coupled to the buffer.

Abstract: An echo canceler has its otherwise conventional tapped delay line basis functions augmented with at least one pair of basis functions realizing two quadrature responses of a complex pole. The frequency of this complex pole is selected as substantially equal to the frequency of the dominant pole of a highpass filter inserted into the echo path of the echo canceler, where the dominant pole is the pole closest to the unit circle. This echo canceler arrangement results in a substantial improvement in attainable echo-return-loss enhancement (ERLE).

Abstract: A method and system for improving performance of an echo canceller with low additional complexity. Specifically, the present invention discloses using information from internal variables of Code Excited Linear Prediction (CELP) based codecs in a digital communication network to significantly improve the rate of convergence of the echo canceller. In one embodiment of the present invention, an error signal associated with a voice signal is filtered by a transversal filter using Linear Predictive Coefficients (LPC) coefficients to provide filter transfer functions for a Filtered-X Least Mean Squares algorithm. Additionally, an adaptive filter applies the Filtered-X Least Mean Squares algorithm using the pre-filtered voice signal available in a CELP-based decoder to create a synthetic echo signal which is subtracted from the echo signal for attenuation.

Abstract: An echo canceller and a method of echo cancellation using an NLMS algorithm. The NLMS algorithm alternates adaptation between a slow mode, aggressive mode, and inhibition mode, according to the status of the echo canceller. This implementation requires no correlation information. This implementation is simple and not very time consuming. During initial conversation, the NLMS algorithm is in an aggressive mode to ensure fast convergence. After convergence, the NLMS algorithm switches to a slow mode for lower residual echo return. Whenever an abrupt increase of residual echo caused by either double talk or echo path change is detected, the current adaptive filter coefficients are retained and then updated in the aggressive mode. The decision as to which echo estimation (the response of the retained filter or the response of the aggressively adapted filter) is subtracted from the reference echo for echo cancellation within a short detection delay period is determined according to the Geigel algorithm.

Abstract: A method of echo cancellation for a signal transmission system in which the size of the step used to adapt filter coefficients is adjusted in accordance with the echo delay and in which stationary signals are avoided by determining when the mean and variance of coefficients obtained from a second order linear predictive coding analysis of successive far end samples exceed preset thresholds.

Abstract: An echo canceling apparatus includes a long-response echo canceler that removes an echo of an incoming signal from an outgoing signal by operating on a most recent segment of the incoming signal, stored in a tapped delay unit. The apparatus has a control unit that calculates a cancellation characteristic for each tap of the tapped delay unit. The cancellation characteristics are used to select a subset of taps corresponding to a changing component of the echo, the subset being empty if there is no changing component. A short-response echo canceler, operating on the outputs of the selected taps, cancels the changing component of the echo left uncanceled by the long-response echo canceler. Alternatively, the step gain of the selected taps is increased. The short-response echo canceler is preferably disabled in the double-talk state.

Abstract: Echoes caused by discontinuities in a telecommunications transmission lines cause double talk which cause divergence problems with echo cancellers. Divergence problems in echo cancellers cause considerable problems in interactive caller response systems, they cause less problems if all parties to a call are human. One of the biggest problems is the effect of the echoes on automatic speech recognition systems. The echo canceller method and system presented cancels echoes to reduce their interference with ASR and yet provides stability in the presence of double talk.

Abstract: Processing complexity in an acoustic echo canceller can be reduced by partitioning the input near-end and far-end signals into respective pluralities of frequency subband signals. Also, due to the complex conjugate symmetry of conventional Discrete Fourier Transform matrices, not all of the subband signals of each plurality need be subjected to echo cancellation processing. The subband signals that would have otherwise been produced by echo cancellation processing are instead provided by producing complex conjugates of selected ones of the subband signals that have been subjected to echo cancellation processing.

Abstract: An adaptive filter suitable for network echo cancellation and other applications contains a coefficient vector update device for feeding coefficient vector updates to a finite impulse response filter in accordance with fast converging algorithms. A double talk detector is included for causing filter adaptation to cease in the presence of double talk in the system being echo cancelled. The coefficient vector update device utilizes a proportional affine projection algorithm to provide fast convergence of the filter system and improved performance over other filter devices utilizing different fast converging algorithms.

Abstract: A speaker-phone that is operable to perform system change detection, double-talk detection and if desired, optimal step-size control. The invention differentiates between system change and double-talk and is operable to these two situations differently. The invention does not rely solely on the energy measurements that are employed by conventional speaker-phones. Highly reliable system change detection is offered when compared to conventional speaker-phones. The direction and the amplitude of the system change is measured with very good detection probability in the invention. For double-talk detection, the invention does not rely on assumptions that the far-end echoes are of lower amplitude that the near-end signal in performing double-talk detection. The invention performs double-talk detection within a large diversity of echo-paths and background noise levels. A designer may tune the speaker-phone system to ensure proper double-talk detection near low levels of the near-end signal.

Abstract: An active noise suppression system for use in noisy environments includes a dual microphone noise suppression system in which the echo between the two microphones is substantially canceled or suppressed. Noise is cancelled by the use of first and second line echo cancellers, which model the delay and transmission characteristics of the acoustic path between the two microphones. In a first embodiment, a noise suppression system acts as an ear protector, canceling substantially all or most of the noise striking the dual microphones of the ear set. In a second embodiment, a noise suppression system in accordance with the present invention acts a noise suppression communication system, suppressing background noise while allowing speech signals to be heard by the wearer.

Abstract: Handsfree telephony continues to be an increasingly desirable feature of modern telecommunications, whether in a conference room or mobile setting. Fundamental to the user acceptability of these systems is the performance of algorithms for acoustic echo cancellation, the purpose of which are to prevent the far-end signal from being transmitted back to the far-end talker. Most speech coding algorithms are based on some variant of Linear Predictive Coding (LPC), which reduces the amount of bits sent across a channel. Instead of doing echo cancellation in the time domain, the invention involves operating an acoustic echo canceller on the LPC parameters at the receiver, before the decoding stage.

Abstract: The step sizes for tap positions of an adaptive filter (124) in an echo canceller (125) for canceling an echo signal are individually selected based upon a profile of the echo environment. A training signal (112) is injected into the echo canceller and to the echo path during a training interval, or period. The adaptive filter profiles the echo path during the training interval. Respective step sizes &agr;[k] for the tap positions are generated as a function of the echo profile so generated.

Abstract: An acoustic transmission is retransmitted by telephone to a network operations center over one path. The network operations center receives acoustic transmissions containing inaudible marks from other sources over secondary paths. The retransmitted acoustic transmission and secondary transmissions are processed and compared to determine a best match between the retransmitted acoustic transmission and one of the secondary transmissions. An acoustic mark in the best match secondary transmission is decoded and associated with the retransmitted acoustic transmission. This association in combination with decision rules stored at the operations center may be used to trigger a variety of transactions with the user.

Abstract: An echo canceler provided with a noise detecting section which holds signals showing a presupposed ambient noise such as a horn of a vehicle. The noise detecting section is adapted to monitor a sending signal received through a microphone and, while detecting such ambient noises in the sending signal, an adaptive filter outputs a pseudo echo existing immediately before the ambient noise is detected. Even during the generation of ambient noises, an echo noise can be eliminated and the generation of a foreign voice (noise) can be avoided as well.

Abstract: Switch 141 continuously switches between received signal 2 and a supplemental signal, which is obtained by processing the received signal 2 through filter 145 to use the output of switch 141 in place of received signal 2. Accordingly, adaptive filters 122 and 124 operate sometimes by using received signal 2 as the input signal and sometimes by using the supplemental signal as the input signal, so that it is possible to obtain adaptive filter coefficients by using twice the number of conditional equations as the case of using only received signal 2 as the input signal. Therefore, since the adaptive filter coefficients do not becomes indefinite, it is possible to converge the coefficients to the correct values. Further, since switching period between the original and the supplemental signals is controlled to be longer than the sampling period of the received signal, it is possible to suppress aliasing distortion of the received signal directly supplied to a speaker and to be maintain better sound quality.

Abstract: When an echo attenuation amount ACOM(i) is smaller than a desired echo attenuation amount ACOM, and an echo cancellation amount ACANC(i) is equal to or greater than a echo/noise ratio E/N(i), a block length computer increases the block length BL(i) and outputs that increased block length BL(i) to a block length controller. The echo canceler comprises such a block length computer and such a block length controller, wherefore the echo attenuation amount can ultimately be made to be equal to or greater than the desired echo attenuation amount.

Abstract: A zero-delay structure to be used in sub-band echo cancellation systems. In telecommunications, a delay is usually introduced in sub-band decomposition and/or sub-band reconstruction. This invention will help remove this. This structure not only removes the delay but does not sacrifice the performance of the echo canceller. It is to be used in combination with any sub-band-based echo canceller or wavelet echo canceller.

Abstract: The present invention provides an adaptive filter and a learning method therefor to eliminate real output signals generated by input signals and an environment signal response. The adaptive filter receives the input signals from an input signal source and generates estimated output signals according to an filter coefficients and the input signals. The first step of the learning method is storing the input signals and the real output signals within a past time period. Then, a predictive input signal and a predictive output signal are generated according to the input signals and the real output signals in the memories, respectively. Finally, the filter coefficients is updated according to the predictive input signal, the predictive output signal, one of the input, signals and one of the real output signals, causing the estimated signals to approximate the real output signals.

Abstract: The invention is a method and apparatus for performing echo cancellation utilizing a Proportionate Normalized Least Mean Squares (PNLMS) algorithm having a high convergence rate and low complexity. The invention utilizes a proportional step-size adaptive algorithm which utilizes the signs of the regressor rather than the regressor value itself in the gradient calculation.

Abstract: A method of canceling echo signals in a telephone network operating in a double talk mode has been developed. A system for implementing the method is also presented. In the invention the conventional normalized least mean square (NLMS) algorithm for echo cancellers is modified such that echo path changes continue to be tracked even after a double talk condition has been detected. The modified algorithm, known herein as a double-talk normalized least mean square (DNLMS) algorithm adaptively adjusts the convergence rate based on the power difference between the echo signal or the residual signal and the far end signal.